Are AI-Generated Images Legal to Sell on Etsy or Shutterstock in 2026

The answer splits cleanly by platform — and one of the biggest stock photo marketplaces has already made its position permanent. Shutterstock explicitly rejects all contributor AI-generated submissions. Etsy, meanwhile, updated its policy in June 2025 to allow AI art with mandatory disclosure. Same product category, completely opposite regulatory posture. That gap defines the landscape sellers need to navigate right now.

Why does this matter in September 2026? The Supreme Court declined to hear Thaler v. Perlmutter in March 2026, reinforcing that purely AI-generated images hold no copyright protection in the U.S. That ruling didn’t kill the market — it reshaped it. According to AutoKeyWorder’s 2026 platform analysis, one contributor earned $3,976 in their first year selling AI images on Adobe Stock alone. The commercial opportunity is real. The legal minefield is equally real.

Key Takeaways

• The Supreme Court’s March 2026 decision in Thaler v. Perlmutter confirmed that purely AI-generated images cannot be copyrighted in the U.S. — but selling them commercially remains legal.
• Shutterstock bans all contributor AI submissions; Adobe Stock accepts them with mandatory disclosure and currently has 47.85% of its library classified as AI-generated.
• Etsy permits AI art sales following its June 2025 policy update, but requires explicit disclosure in listing descriptions and prohibits “Handmade” classification.
• Three hard legal risks apply across all platforms: generating likenesses of real identifiable people, recreating IP-protected characters, and failing to disclose AI involvement where required.
• Adding significant human creative editing to AI outputs — Photoshop compositing, custom color grading, original layout decisions — strengthens copyright eligibility under current U.S. Copyright Office guidance.

How the Legal Ground Shifted in 2026

The copyright question has been building since 2022, but 2026 brought two concrete developments that changed seller calculus.

First: the Supreme Court’s March 2026 decision. By declining to hear Thaler v. Perlmutter, the Court left the U.S. Copyright Office’s position intact — prompts alone don’t establish copyright. Two users who generate images from identical prompts both hold commercial usage rights to their outputs, but neither holds exclusive copyright. According to AutoKeyWorder, 51 active copyright lawsuits currently target AI companies, with the Andersen v. Stability AI class action heading to trial in September 2026. The training-data question isn’t settled. The output-ownership question largely is.

Second: platform divergence accelerated. Adobe Stock now receives 29+ million AI images monthly and has classified 47.85% of its library as AI-generated. Shutterstock went the opposite direction — blanket rejection. This isn’t a temporary policy gap. It reflects fundamentally different business models. Adobe generates revenue from AI tools like Firefly, so embracing AI content reinforces that ecosystem. Shutterstock’s traditional contributor base and enterprise licensing clients pushed hard for differentiation from AI-saturated stock libraries.

The FTC also increased pressure on commercial AI disclosure in 2026. Across most U.S. jurisdictions, disclosure has shifted from “best practice” to legal expectation in commercial contexts. Sellers ignoring this aren’t just risking platform bans — they’re building potential FTC exposure.

Platform-by-Platform Reality Check

Etsy: Open, But Conditional

Etsy’s June 2025 policy update was specific. AI art is permitted. Disclosure in the listing description is mandatory. The item must never be classified as “Handmade.” Purchased prompt packs violate policy outright — Etsy distinguishes between AI art you created through your own creative process and repackaged prompts someone else developed. According to ZSky AI’s 2026 legal framework guide, many buyers on Etsy actively search for AI art, so disclosure doesn’t hurt discoverability. Digital downloads typically price between $3–$25.

Adobe Stock vs. Shutterstock: The Real Split

The Shutterstock rejection is firm. No disclosure workaround exists, no submission category for AI — the ban applies to contributor uploads regardless of how much human editing was added post-generation.

Adobe Stock’s market position is almost the inverse. New accounts face an upload cap of roughly 51 images per week, which limits early momentum but doesn’t fundamentally block the channel. Metadata quality — titles, keywords, categories — determines ranking and discoverability more than image volume, according to AutoKeyWorder’s platform analysis.

Print-on-Demand Platforms

Redbubble accepts AI with a disclosure checkbox. New accounts start at 5 AI uploads per day, scaling to 50/day at meaningful sales volumes. Per-sale earnings run $1–$5. Displate accepts AI with no disclosure field, but reports a 200,000+ item approval backlog. TeePublic offers a $4 standard royalty — dropping to $2 during sales — with no formal AI disclosure field.

The Three Hard Legal Lines

Regardless of platform, three categories create genuine legal risk.

Recognizable individuals. Generating a realistic likeness of a living person for commercial use without their consent violates right-of-publicity laws across most U.S. states. AI didn’t change that law — it just made creating such images easier and violations more frequent.

IP-protected characters and trademarks. Disney, Marvel, Nintendo — it doesn’t matter if the image was AI-generated. Trademark infringement applies the same way it would to a human artist. Accidentally including brand logos in backgrounds creates liability too. According to Louplr’s 2026 commercial use guide, inspecting images at 100% zoom before listing — specifically to catch unintended brand elements — is now a standard compliance step among serious sellers.

Style-of named living artists. Prompting “in the style of [living artist]” sits in ethical and legal gray territory. The Andersen v. Stability AI trial running in September 2026 will likely sharpen this boundary. Sellers who’ve built libraries around specific artist style prompts should watch that case closely — the verdict could land before Q1 2027 and reshape training-data liability across the industry.

Practical Steps for Sellers Right Now

The core problem is a fragmented compliance landscape with no single rulebook. Platform rules, federal copyright law, FTC disclosure expectations, and state right-of-publicity statutes all pull in different directions. That’s not a reason to avoid the market — it’s a reason to build a clean process from day one.

Building a stock portfolio. Adobe Stock is currently the clearest path for AI sellers. Set up disclosure correctly from account creation, invest in metadata quality — keyword research, accurate categorization — and expect roughly 3–6 months before meaningful earnings materialize given the 51-upload-per-week cap on new accounts. The volume ceiling is real, but sellers who clear it consistently report compounding returns as catalog size grows.

Selling digital art on Etsy. Draft disclosure language before listing anything. Something like: “This design was created using AI image generation with my creative direction, including prompt development and post-processing.” Document your generation dates and prompts. Never classify as Handmade regardless of how much editing you add — that classification triggers policy violations even when the human contribution is substantial.

This approach can fail when sellers treat disclosure as a one-time checkbox rather than a consistent practice. Platform audits do happen, and retroactive compliance across hundreds of listings is far harder than building it in upfront.

Protecting earnings long-term. Add meaningful human editing — compositing, color correction, layout decisions — to strengthen copyright eligibility. According to ZSky AI’s framework, commercial usage rights granted by AI platforms and copyright ownership are separate legal concepts. You can sell legally without owning traditional copyright, but ownership matters if someone copies your specific output and you want legal recourse.

What Comes Next

The legal framework for selling AI art in 2026 is functional but incomplete. Four points summarize where things stand:

• Selling AI images commercially is legal across most major platforms, with Shutterstock as the prominent exception.
• Copyright protection requires demonstrable human creative input beyond prompt writing.
• Disclosure is no longer optional — it’s moving toward legal requirement at the federal and platform level simultaneously.
• The Andersen v. Stability AI verdict will be the most significant legal event for this market in the next 12 months.

Near-term, expect more platforms to formalize AI disclosure fields as FTC pressure increases — TeePublic and Displate are the obvious candidates. If Andersen produces a broad ruling on training-data liability, AI image tool pricing and terms of service will likely shift, potentially affecting what commercial rights platforms grant to sellers downstream.

The one clear action available right now: document everything. Generation dates, prompt records, post-processing steps. That documentation is what separates sellers with defensible copyright claims from sellers with none — and in a market this unsettled, that difference matters more than most people realize.

The platforms have picked sides. The job now is choosing the right one for your model and building compliantly from the start.

Sources: ZSky AI | AutoKeyWorder | Louplr

References

1. What is Etsy’s stance on AI creations?
2. Can You Sell AI Generated Images? Copyright Rules Explained (2026)
3. Can You Sell AI Art on Etsy? 2026 Policy Guide | XHBT®

Photo by Steve A Johnson on Unsplash